Intel Core Ultra 9 285K review: surprising performance from a non-hyperthreaded CPU
Has Intel dropped the ball with its latest Core Ultra 9 flagship?
WePC is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Prices subject to change. Learn more
Intel has officially released its new rebranded Core Ultra CPUs, but if you’re in the habit of denying change, these are the brand’s latest 15th-generation CPUs. The Core Ultra 9 285K sits at the top as the flagship for this generation, so we’re expecting big things, despite some bold changes on Intel’s part.
The Core Ultra 285K features 24 cores and 24 threads, meaning, for the first time in over 22 years, the brand has moved away from hyperthreading – a change that continues to baffle us. Intel’s big.LITTLE core architecture remains present, with an 8 P-core and 16 E-core configuration – similar to the 14900K. But how will the new Core Ultra 9 285K stack up against both its predecessor and its main AMD rival, the 9950X?
- Core configuration: 24 (8+16)
- Threads: 24
- P-Core Speed: 5.5 GHz (boost) 3.7 GHz (base)
- E-Core Speed: 4.6 GHz (boost) 3.2 GHz (base)
- DDR5 support: DDR5 @ 6400 MT/s
- TDP / PL1 / PL2 / PL2 (extreme): 125W / 250W / 250W / 295W
We expected the Core Ultra 9 285K to be the CPU to continue the commanding lead that Intel had in single core performance and blow the latest AMD CPUs out of the water. With a die shrink of 7 nm, it seems we expected too much from the Core Ultra series flagship. It does edge out in synthetic workloads and multi-core performance, but it barely keeps up in gaming. Much more efficient though, if that’s your thing. For productivity workloads, however, this CPU is a beast that cannot be tamed.
- Very efficient across all workloads
- Easier to cool than most
- Cheaper than AMD flagship
- Much better multi-core performance than 9950X
- Underwhelming gaming performance
- Feels like something is missing
- Too focused on efficiency
Let’s take a deeper dive into the specifications and see what’s changed regarding the 15th gen, Core Ultra 9 flagship
Specifications and comparison
The Core Ultra 9 285K is this generation’s flagship processor so it should feature the best specifications that Intel offers. If you’re interested and need to know all of the CPU specifications quickly, we have listed them below – alongside 14900K comparisons so you can see generational changes.
| Component | Core Ultra 285K | 14900K |
|---|---|---|
| Cores | 24 | 24 |
| Threads | 24 | 32 |
| Hybrid architecture | P-core: 8 E-core: 16 | P-core: 8 E-core: 16 |
| Base frequency | P-core: 3.7 GHz E-core: 3.2 GHz | P-core: 3.2 GHz E-core: 2.4 GHz |
| Boost frequency | P-core: 5.5 GHz E-core: 4.6 GHz | P-core: 5.6 GHz E-core: 4.4 GHz |
| Boost 3.0 | 5.6 GHz | 5.8 GHz |
| Turbo clock | up to 5.7 GHz | up to 6 GHz |
| Memory speed | 6400 MT/s | 5600 MT/s |
| PCI Express lanes | 20 (Gen 5) | 16 (Gen 5) |
| L2 cache | 3 MB (per core) | 2 MB (per core) |
| L3 cache | 36 MB (shared) | 36 MB (shared) |
| TDP / PL1 / PL2 / PL2 (extreme) | 125W / 250W / 250W / 295W | 125W / 253W / 253W / Unlimited |
| Process size | 3 nm | 10 nm |
| Socket | LGA 1815 (compatible with Z980) | LGA 1700 (compatible with Z790, B760, Z690, B660, H610) |
We observed many differences against the previous generation of Intel flagship, the biggest of which is the process size being reduced from 10 nm to 3 nm – allowing more transistors to be squeezed into the same footprint, increasing IPC.
Almost everything has been improved across the board, apart from core frequencies – Intel has decided to turn down the clocks this time rather than boost them. This could be for the sake of energy efficiency, as the extreme PL2 limit is now limited to 295W rather than having no limit, which can be manually lifted in the BIOS should you want to.
Additionally, more PCIe lanes add compatibility and flexibility when running additional PCIe devices without interfering with the GPU. This could also hint that, in the future, PCIe slots could be switching to x20 over the current x16, resulting in increased bandwidth. But we don’t know anything about that yet. Either way, it’s a future-proofing measure. AMD’s Ryzen 9000 series already has up to 24 Lanes.
Design and Gallery
The 285K’s design has not changed too much physically over the previous generation, but the pin configuration switching to 1851 (as the socket has now changed to LGA 1851) means more contact pads on the bottom of the CPU. Internally, however, things are very different.
Intel has decided to ditch hyperthreading, resulting in some less-than-ideal performance metrics. This is likely an efficiency play on Intel’s part, as applications that use single-threaded asynchronous (non-blocking) I/O are more efficient on single-threaded CPUs. In Windows, an async/await protocol makes single-core implementation easy. In theory, this should make the Core Ultra CPUs work a little faster in the more popular operating system.
If you are interested enough, you can read this thread on Stack Overflow for a better explanation – greater minds than mine have laid out more encompassing explanations of this everything-but-easy topic.
Performance
Before we dive into the ins and outs of how this CPU performed in our benchmarks, we must look at the components we used to test the 285K. This helps us understand how and why the CPU scored how it did – here are the components that comprise the WePC testing rig:
| Component | WePC test rig |
|---|---|
| CPU | Intel Core Ultra 285K |
| Cooler | Corsair H150i Elite LCD |
| Motherboard | ASUS Maximus Extreme Z890 |
| Memory | Corsair Dominator Platinum DDR5 @6800 MHz |
| GPU | MSI RTX 4070 Ti |
| PSU | ASUS Thor Platinum II 1000W |
| Case | Cooler Master Masterframe |
As you can see, we spare no expense when testing the latest hardware because we want to give every component the best chance to give its all. With that in mind, here’s how the 285K performed in our gaming benchmarks.
Graphs
To make things easier to digest, we created some graphs to display all of the important benchmarking information.
285K vs 9950X review synthetic multi core
285K vs 9950X review average power consumption
285K vs 9950X review power
285K vs 9950X review synthetic single core
Gaming benchmarks
We tested the Core Ultra 285K in a few games to test a broad range of capabilities. All of the games were tested in 1080p low settings, to avoid being GPU-bound since we are testing the CPU, not the GPU.
These games ranged from the more CPU intensive to the easier to run and are listed in order of benchmark:
| Metric | CS2 | Days Gone | CP 2077 | Doom Eternal | Horizon FW | SOTF | Frost punk 2 |
|---|---|---|---|---|---|---|---|
| Score (FPS) | AVG: 313.8 99%: 133.7 | AVG: 236.9 99%: 153.4 | AVG: 209.7 99%: 149.1 | AVG: 499.8 99%: 393.8 | AVG: 202.6 99%: 153.1 | AVG: 185.1 99%: 147.1 | AVG: 231.1 99%: 168.1 |
| AVG temp (package) °C | 44 | 46 | 52 | 51 | 51 | 44 | 46 |
| Max temp (package) °C | 48 | 51 | 65 | 54 | 55 | 46 | 51 |
| Average PPT (W) | 48.5W | 61.88W | 110.25W | 108.04W | 91.54W | 55.47W | 71.32W |
As you can see, the 285K scores well in most gaming benchmarks while keeping power levels relatively low. Gaming was never as CPU-intensive as productivity tasks, so it’s not uncommon to see a CPU below 120W when gaming. The anomaly to all these results has to be Cyberpunk 2077 – no matter how hard we tried or what settings we used, we never managed to get the game to benchmark on par with other tests.
However, when we compared the 9950X, the scores looked less impressive. AMD’s newest flagship manages to edge ahead in most cases against the 285K. Granted, we were on the base BIOS and pre-release drivers, so this could change post-release, but this is all we have for now.
| CPU | CS2 | Days Gone | Doom Eternal |
|---|---|---|---|
| Core Ultra 9 285K | AVG: 313.8 99%: 133.7 | AVG: 236.952 99%: 153.415 | AVG: 499.82 99%: 393.848 |
| Ryzen 9 9950X | AVG: 323.7 99%: 136.43 | AVG: 237.4 99%: 133.32 | AVG: 507.658 99%: 386.563 |
It’s a fairly close comparison, but the 9950X wins by a hair every time. It’s so close that I’m almost tempted to say that these are the same, and it’s just a margin of error – but I don’t want to take away from the 9950X. That said, these results do not bode well for Intel, especially with AMD’s new X3D series looming large. The 9800X3D has been confirmed to be released on November 7th, and will likely offer massive gains over the 9950X in gaming workloads.
If we look at the last generation, the 7800X3D offered a 20% average uplift in gaming performance compared to the 7950X, so if we see anything even remotely close to that with the 9000 series, it’s going to blow Intel out of the water.
Synthetic benchmarks
| Metric | CPU Z | Cinebench R23 | Cinebench 2024 | Geekbench | Blender render | 7 Zip Comp (10 passes) |
|---|---|---|---|---|---|---|
| Score (points) | Single: 896 Multi: 18,866 | Single: 2,134 Multi: 42,399 | Single: 132 Multi: 2,441 | Single: 3,166 Multi: 23,541 | Monster: 269.9 Junkshop: 179.8 Classroom: 131.9 | 43.84 seconds |
| AVG temp (package) °C | 54 | Single: 45 Multi: 62 | Single: 46 Multi: 69 | 41 | 66 | 61 |
| Max temp (package) °C | 67 | Single: 51 Multi: 73 | Single: 55 Multi: 75 | 65 | 75 | 69 |
| Average PPT (W) | 102.25W | Single: 29.98W Multi: 160.06W | Single: 34.5W Multi: 203.2W | 26.64W | 190.3W | 149.06W |
It was a good showing from the Core Ultra 285K in terms of synthetic benchmarks, beating out its 9000 series competitor by a wide margin (when it’s not using PBO). The 9950X scored 40,166 points in the CB multi benchmark, and the 285K brought home a score of 42,399. This is a massive achievement considering the 9950X has 8 more threads.
| CPU | Score (CB Multi) |
|---|---|
| Core Ultra 9 285K | 42,399 |
| Ryzen 9 9950X | 40,166 |
However, the Ryzen 9 9950X manages to edge ahead of the 285K when PBO is enabled, although it has to work so hard that it throttles to achieve that score, even with a 420mm AIO cooler. The Core Ultra 9 feels more focused on efficiency rather than absolute no-holds-barred performance. There seems to be a lot of that going around lately – you couldn’t exactly call it well-received when AMD attempted it with the 9000 series.
When looking at the multi-core scores specifically, it’s hard to believe that this CPU has fewer threads than the 9950X with the level of performance it outputs. In Geekbench specifically, the 9950X managed to bring in 20,000 points, but the 285K scores over 23.5K in the same test.
Temperature and efficiency
Compared to the 9950X, the 285K is much more efficient and generally cooler. Yeah, you read that right, an Intel CPU is cooler and more efficient than an AMD CPU – a plus for the energy-conscious among you. We also ran the 285K with just the efficiency cores enabled, since Intel boasts a massive 32% performance uplift on its Skymont E-cores.
| CPU | Cinebench R32 (multi) | CS2 | Days Gone | Doom Eternal |
|---|---|---|---|---|
| Core Ultra 285K | Score: 42,399 Max temp: 73°C Max PPT: 160.06W | AVG: 313.8 Max temp: 48°C Max PPT: 48.5W | AVG: 236.952 Max temp: 51°C Max PPT: 61.88W | AVG: 499.82 Max temp: 54°C Max PPT: 108.04W |
| Core Ultra 285K (E-core only) | Score: 25,394 Max temp: 57°C Max PPT: 85.93W | AVG: 146.27 Max temp: 47 Max PPT: 28.4W | AVG: 216.727 Max temp: 45°C Max PPT: 40.38W | AVG: 510.843 Max temp: 49°C Max PPT: 67.71W |
| Ryzen 9 9950X | Score: 40,166 Max temp: 64°C Max PPT: 167.67W | AVG: 323.7 Max temp: 62°C Max PPT: 113.55W | AVG: 237.4 Max temp: 73°C Max PPT: 108.33W | AVG: 507.658 Max temp: 62°C Max PPT: 155.96W |
As you can see, the 285K stayed much cooler than the 9950X and used less energy in some cases. The testing with purely E-cores had to involve one P-core for the system to boot properly – likely how it managed to do so well in some gaming workloads despite being massively handicapped. We’re not recommending that you do this, we just wanted to compare for fun.
The interesting thing about having to enable one P-core for the 285K to work, is that this wasn’t the case. On the Alder Lake processors when the big.LITTLE core architecture was introduced, you didn’t need to have any P-cores enabled on the 12900K. The gaming performance certainly draws from this lone active P-core, though it seems that in CS2, the system decided to use an E-core. Great job task scheduler.
There was only one case where the 9950X was cooler than the 285K and that was in the Cinebench multi-core tests, understandable given that the cores are not hyperthreaded and have to work at maximum capacity to have a shot at scoring near the top spot – the hard work paid off as it managed to edge out the 9950X by over 2,000 points.
Price
The Core Ultra 9 285K has launched at $589, on par with the last generation’s flagship, the 14900K. We’re happy to say that 285K beats the 14900K in all of the benchmarks performed, so it (at the very least) improves upon the last generation of processors – offering a better value. However, you do have to factor in the cost of a shiny new motherboard on the LGA 1851 standard.
The AMD 9950X retailed for $649 but you can sometimes find it on sale for $599, just $10 more than the 285K (assuming it stays at retail) for the same gaming performance, which doesn’t seem worth it – and you’re better off getting the 285K if you need more productivity performance.
Is the Intel Core Ultra 9 285K worth it?
The Core Ultra 9 285K is a fantastic selection for the best productivity performance possible. You could also argue that this CPU is worth it for mixed-workload machines, such as gaming and productivity. However, if you want to build a pure gaming machine we advise you to head away from Intel and look to AMD X3D, specifically the 9800X3D when that is released. We predict that the 9800X3D will smoke the 285K in gaming performance.
Final word
Intel Core Ultra 9 285K
